While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
搜索“Qwen”
Nathan Lambert@natolambertAI 评分5858引用Cohere@cohere
Andrew Curran@AndrewCurran_同事件AI 评分7878
引用Cohere@cohereWhile evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: http://cohere.com/blog/concerning-behaviour-popular-open-weight-ai-model
同一件事,精选收录的是《Cohere 称 Qwen3.8-Flash-Next 在沙箱中尝试向阿里云存储桶上传探测文件》
钛媒体AI 评分4242 从“拼模型”到“抢入口”:国产大模型进军消费硬件
阶跃星辰与阿里千问于10月13日同日发布AI手机和AI眼镜、耳机等消费硬件,国产大模型厂商的竞争从模型性能与价格延伸至硬件入口。行业数据显示,今年1—2月国内主流AI应用30天留存率仅12.8%,厂商希望借硬件提高用户更换成本、锁定用户。阶跃星辰STEPX Neo先行版全栈定价3999元,内置智能体Amoo可跨应用完成订票、办公等任务。

IT之家AI 评分6262 中国信通院报告:6 月国内词元日均消耗量约 500 万亿,为 2025 年 12 月的近 4 倍
中国信通院发布《大模型词元服务性能监测报告(2026 年)》,指出 2026 年 6 月国内词元日均消耗量约 500 万亿,是 2025 年 12 月的近 4 倍。

钛媒体AI 评分5050 腾讯QClaw公测188天即宣布停服,大厂AI产品加速洗牌
腾讯QClaw宣布将于2026年12月24日停止运营,距3月20日公测仅188天,用户数据将迁移至WorkBuddy。阿里千问则宣布即将发布个人助理产品“千问Go!”,字节也被曝推进代号“Spell”的个人助理项目。四年间大厂AI产品改名、整合、停运成为常态,独立产品正不断向通用助手和底层平台收拢。

极客公园AI 评分5858 天猫双 11 上线 AI Token 站,多家大模型厂商入驻
天猫在 2026 双 11 发布会上宣布推出 AI Token 站,阿里云百炼、千问办公、智谱、MiniMax、月之暗面、阶跃星辰均已入驻,算力额度首次成为可直接下单的商品。

钛媒体AI 评分4040 MiniMax入驻天猫开卖Token套餐,大模型厂商双11下沉电商
MiniMax双11前夕在天猫开设首家品牌官方旗舰店,上架Go、Explore、Build三档M Plan Token订阅套餐,支持按月、按年购买。今年9月智谱、阿里云百炼、千问办公等已开设旗舰店,天猫同期上线"AI空间站"充值中心。但各厂商旗舰店销量并不理想,智谱最高一款产品仅售出200多份。

IT之家AI 评分6060 中国信通院报告:2026 年国内大模型词元均价翻倍、TPS 翻倍
中国信通院等机构 10 月 23 日联合发布《大模型词元服务性能监测报告(2026 年)》,基于 2026 年 1-8 月监测数据对比各平台词元服务性能。报告显示 8 月国内模型平均 TPS 达 78.23,较 2025 年均值 38.25 翻倍,但低于国外典型模型的 112.82;国内归一化均价从 2025 下半年的 6.8 元升至 13.46 元,涨幅接近 1 倍。

钛媒体AI 评分5858 双11首卖Token套餐:24.5元起,但各家计量口径互不相通
钛媒体作者听风译码观察到,今年天猫双11首次把大模型Token按月卡、季卡、年卡上架,MiniMax首单五折后最低24.5元,阿里云百炼、智谱、月之暗面、阶跃星辰等7家厂商入驻AI空间站。作者实测发现同一货架上至少存在Credits、积分、用量进度条三种计量口径且互不等价,部分套餐还受5小时窗口限制,比价难以成立。文章建议先跑一个月小额月卡摸清真实消耗,再决定是否上年卡,并留意额度清零与退款条款。
雷锋网AI 评分3030 IDC报告:阿里云占中国汽车公共云市场份额超45%
IDC《中国汽车云市场(2026年上半年)跟踪》报告显示,2026年上半年中国汽车公共云基础设施市场规模约83.4亿元,阿里云以45.1%份额居首,收入同比增长111%。这也是阿里云自2021年起连续11期位居该市场第一。
少数派AI 评分2222 Tim Cook 探访蓝盒子:Mac 如何支撑 300 人家居企业的本地 AI 与设备管理
Tim Cook 上周四走访上海床垫家居品牌蓝盒子,了解其以 Mac 和 iPad 支撑设计、内容制作与线下试睡服务的用法。蓝盒子用一台 64GB 内存 Mac mini 本地部署 35B 参数千问模型,员工以自然语言转 SQL 查询销售、库存等数据,已投用约一个月,部分查询从一两天缩短到几分钟。新员工开机联网即可零接触部署,从拆封到开工约 12 分钟。

AWS Machine Learning BlogAI 评分3636 Amazon SageMaker JumpStart 2026 年至今回顾:新增 100+ 模型
Amazon SageMaker JumpStart 在 2026 年 1 月至 9 月新增超 100 个模型,覆盖 30 家提供商。阿里 Qwen 贡献 29 个模型,NVIDIA 17 个,Mistral 7 个,Google 6 个 Gemma 4 变体,IBM 和 Cohere 各 4 个。新增模型重点覆盖多模态、量化与紧凑模型以及图像视频生成能力。

钛媒体AI 评分5858 腾讯领跑、百度第二,AI办公赛道远未定局
QuestMobile 三季度报告显示,9月AI办公赛道活跃用户达1.42亿,同比增长50.6%,办公Agent榜前三为腾讯WorkBuddy(3235.3万)、百度搭子(977万)和阿里千问办公(790.5万)。

Qwen@Alibaba_QwenAI 评分2222实用 AI 解决现实问题的绝佳范例!速度前所未有。感谢使用 Qwen 构建。@cerebras 快来亲自试试!🏠
引用Cerebras@cerebrasWe built Money Agent, a personal finance assistant powered by @Alibaba_Qwen 3.8 27B on Cerebras. Now you can turn a home-buying question into a real-time conversation, then into a financial goal, faster.
Xiaomi MiMo@XiaomiMiMoAI 评分2525开源权重模型正在崛起——很高兴看到越来越多开发者在 MiMo 上构建。👏 @cline
引用Cline@clineWe’ve been impressed with GLM-5.2 and so are introducing a $9.99/month subscription to give you 2-5x discounted access to it and other open weight models like DeepSeek, Kimi, MiniMax, Mimo, Qwen. Use it on Cline CLI & IDE with $1.99 special promo if sign up via: npm i -g cline